Why did some of your people thrive with AI and others didn’t?

Two people on your team got the same tool on the same day.

Same onboarding. Same access. Same demo from the same vendor.

Six weeks later, one of them has restructured how they run their entire workflow around it. The other still opens it occasionally, types something tentative, closes it. You’ve watched this happen. You’ve probably told yourself it’s a learning curve. A personality thing. Maybe a generational gap.

Keep sitting with that explanation. Because something about it feels off.

The tool didn’t change between those two people. The prompts weren’t harder for one than the other. The interface is the same interface. So when the same rollout produces wildly different outcomes across a team, the instinct is to look at the tool. Adjust the training. Run another workshop. And sometimes that helps, at the margins.

But the stabilize/destabilize split — the one where some people expand into AI and others contract around it — that split was already there. The tool just made it visible.

Here’s what’s actually happening. AI behaves like secure attachment. It’s available when you need it. It doesn’t punish you for asking a bad question. It doesn’t sigh when you circle back to something you already covered. It doesn’t have a mood. It doesn’t remember that you were short with it last Tuesday. For people who’ve spent years in environments where asking for help was a liability, where showing uncertainty meant losing ground, where the wrong question in the wrong meeting cost you something — AI feels like relief. They expand. They experiment. They come back.

For people who’ve learned to perform certainty, to protect their expertise as a form of status, to stay in control of what others see — AI is destabilizing. Every interaction is a small exposure. Every prompt is a record of what they didn’t know. They contract. They use it minimally. They find reasons it doesn’t quite work for their role.

Same tool. Completely opposite responses. The answer isn’t in the tool.

We brought this with us.

The relational dynamics underneath uneven AI adoption — the fear, the performance, the hunger for safety, the protection of status — none of that arrived with the software. It was already running in your team. In your meetings. In who speaks up and who doesn’t. In who asks for help and who disappears when they’re stuck. It just gave those patterns a new surface to show up on.

And that’s the part that’s hard to sit with. Because if the problem is emotional reality, not technology, then another rollout won’t fix it. A better prompt library won’t fix it. The gap you’re watching isn’t a capability gap.

Psychological safety. Permission to not know. Room to experiment. Trust that failure won’t cost them their standing. These aren’t soft skills. They’re the actual infrastructure your team runs on. AI just made the infrastructure visible.

So now the harder question.

You’re watching two people respond to the same tool in opposite ways. You’ve been thinking about what they need to change. But the manager is the variable here, not the tool. The environment those two people are operating in — the one where asking questions is safe or isn’t, where uncertainty is allowed or isn’t, where experimentation is rewarded or quietly penalized — that environment has an author.

This is the part where it gets uncomfortable. Because the same mirror that’s reflecting your team’s dynamics is also reflecting yours. The leader who says “my door is always open” but whose reactions have taught people to never walk through it. The one who champions psychological safety in all-hands and then goes cold when someone surfaces a real problem. The one who genuinely doesn’t know that the team has learned to manage their mood before managing their work.

AI is a mirror. It doesn’t lie. It shows you what people bring to it, which shows you what they’ve learned to bring to everything. And what they’ve learned, they learned somewhere.

The question isn’t whether your team is ready for AI.

Psychological safety. Permission to not know. Room to experiment. Trust that failure won’t cost them their standing. These aren’t things you can train into people in a workshop. They’re responses to conditions. They grow or they don’t, depending on what the environment has consistently made possible.

So the real question — the one worth sitting with — is what your team has learned is true about this particular room. What they’ve learned is safe to show. What they’ve learned to hide.

Are you ready to see what the tool is showing you about the room you built?